Senior Applied Scientist, Crunchyroll

• Experience: You bring 5+ years of experience in applied machine learning, recommendation systems, search/ranking, experimentation, or a closely related area, with a track record of driving measurable product impact
• Scientific Depth: You have strong foundations in machine learning, statistics, experimental design, and causal thinking, and you know how to choose the right level of modeling complexity for the problem at hand
• Recommendation Expertise: You have hands-on experience with at least some of the following: collaborative filtering, retrieval and ranking systems, representation learning, sequence / generative models, bandits, graph methods, or personalization for consumer products
• Technical Skills: You are highly proficient in Python and comfortable working with common ML libraries such as PyTorch, TensorFlow, Scikit-learn, XGBoost, or similar tooling. Experience working with SQL, distributed data processing, and cloud-based ML workflows is strongly preferred
• Experimentation Mindset: You know how to design offline and online evaluations, reason carefully about metrics, and connect experimental findings to user and business outcomes
• Cross-Functional Collaboration: You have experience partnering effectively with engineers, product managers, analysts, marketers, and business stakeholders to move from idea to execution
• Communication Skills: You can explain sophisticated modeling decisions and ambiguous findings in a clear, decision-oriented way to diverse audiences
• Educational Background: You hold an MS or PhD in Computer Science, Machine Learning, Statistics, Operations Research, Economics, or a related quantitative discipline, or you bring equivalent applied industry experience

Desirable

• In the role of Senior Applied Scientist for Recommendation and Personalization, you will report to the Director of Data Science and Machine Learning in our Center for Data and Insights
• You will own the research and applied science agenda for personalization, from problem framing and data exploration through model development, evaluation, experimentation, and iteration
• This role is ideal for someone who enjoys combining strong scientific rigor with product thinking to improve user discovery, engagement, retention, and long-term fan value
• You will work across multiple user touchpoints, including app and web interfaces, lifecycle and promotional email campaigns, and flywheels that connect video, ecommerce, manga, and adjacent experiences
• You will help define what great personalization looks like at Crunchyroll, build the evidence to prove impact, and collaborate with engineering partners to ensure the resulting solutions can be productionized effectively
• Lead the research and development of recommendation, ranking, retrieval, and personalization methods tailored to Crunchyroll use cases across streaming, manga, ecommerce, and lifecycle marketing surfaces
• Frame ambiguous business and product questions into clear scientific problems, hypotheses, success metrics, and experimentation plans
• Design and run robust offline evaluation frameworks for recommender systems, including relevance, diversity, novelty, coverage, calibration, and long-term value metrics
• Partner with Product, Analytics, and Engineering to define online experiments, interpret results, and turn learnings into roadmap decisions and model improvements
• Develop user, content, and contextual understanding through feature design, representation learning, segmentation, and behavioral analysis
• Prototype and evaluate a range of approaches, including collaborative filtering, content-based methods, sequence modeling, deep learning, bandits, causal or uplift methods, and LLM-enabled recommendation techniques where appropriate
• Analyze user feedback loops and cross-domain interactions to improve discovery across video, merchandise, manga, and other ecosystem experiences
• Work closely with Machine Learning Engineers to translate promising research into production-ready solutions on our in-house recommendation platform
• Communicate scientific findings, model tradeoffs, and business implications clearly to technical and non-technical stakeholders
• Help establish best practices for experimentation, reproducibility, model governance, and scientific documentation within the personalization and recommendations function